The objective of this core TR&D project is to improve methods for the selective staining of neurons for correlative light and electron microscopy and to develop methods that will help researchers to visualize and quantify neuronal shape and connectivity at EM resolution. Our efforts in this area have been driven by our collaborative efforts with Drs. Charles Wilson, Gordon Arbuthnott and Cali Ingham on obtaining quantitative information on structural changes in striatal dendrites and dendritic spines associated with partial deafferentation. Briefly, we have been using electron microscopic tomography to obtain volume reconstructions of spiny dendrites in both normal neostriatum and in neostriatum in which the dopaminergic input has been lesioned chemically. Spiny dendrites receive the bulk of dopaminergic input within the neostriatum and it was observed at the light microscopic level that loss of this input may lead to both a loss of spines and a change in their morphology. The pilot portion of this project which involved acquiring tomographic reconstructions and quantitative measurements from 20 dendrites has been completed. These 20 dendrites came from 4 animals from both the lesioned and unlesioned neostriata. This population of dendrites yielded length, surface area and volume measurements from over 200 dendritic spines. Two problems were encountered: 1) because of the large variability within each group, we have determined that it will be necessary to obtain additional reconstructions from several more animals per group; 2) it appeared that the spine densities that we were finding were lower than those reported by serial sectioning analysis. Because we were limited to a section thickness of about 2 5m at 400 keV, many spines were likely being cut off of the dendrite at the top and bottom of the section. We performed tomographic reconstructions of spiny dendrites contained within 3-4 5m thick sections from tilt series acquired with the 1meV high voltage microscope in Okasaki, Japan. We found that the spine density was on average 20-30% higher with these thicker specimens, suggesting that indeed, we are underestimating the spine density. To overcome the thickness limitation in calculating spine densities, we are determining the accuracy with which dendritic spines can be counted in Golgi-filled neurons at the light microscopic level by using correlated 3D light and electron microscopic examination. A through-focus series of images was obtained at the light microscopic level of Golgi-impregnated dendrites. Using image deconvolution, a 3D volume was produced and the individual spines segmented and counted. The dendrite was then embedded for electron microscopy, and portions of the dendrites were reconstructed using electron tomography. We have performed spine counts on two correlated volumes thus far and have found 10-20% higher spine densities in the electron microscopic volumes, mostly due to the difficulty in resolving overlapping spines in the light microscope. We are hoping to be able to derive some scaling factors that can be used to correct light microscopic estimates so that these types of analyses c an be performed at the light microscopic level. Another structure that we have been working on is the basket cell "pinceau" in the cerebellar cortex. The pinceau is formed by the axons of basket cells which descend from the molecular layer and completely encase the Purkinje cell axon initial segment and cell body. Despite the impressive size of this axonal ramification and its close association with the Purkinje cell, virtually no synapses are formed on the axonal initial segment. We are performing 3-dimensional reconstructions on this structure to gain some understanding of the relationship of individual axons in the pinceau to the Purkinje cell and to each other. These reconstructions will be used by modelers such as Jim Bower at Cal. Tech. to investigate the possible function of this mysterious and complex structure. We have reconstructed a large extent of the pinceau from 3 Purkinje neurons using relatively low magnification micrographs and have reconstructed several additional basket cell fiber terminations at high er magnification. This work was presented at the Society for Neurosciences meeting this past year. Simultaneous with the structural work, we are investigating the distribution of various macromolecules that are concentrated in the pinceau. We are interested in developing methods for representing the locations of various ion channels (in particular potassium channels), membrane proteins (e.g. PSD-95) and intracellular calcium regulatory proteins in reconstructions in a form useful to modelers and other investigators. We have recently received funding for the creation of a cell-centered database for this type of data.